Most startups announce their seed rounds with a product demo and a waitlist. Agaton announced its seed round with millions in annual recurring revenue already on the books. That's not how stealth mode usually works.

The Stockholm-based AI startup has raised $10 million (roughly EUR 8.4 million) in a seed round co-led by Inception Fund and Alstin Capital, with participation from Seed+speed Ventures, Foundry Ventures, and a roster of angel investors that includes Peter Sarlin (CEO of Silo AI, which sold to AMD for $665 million) and Kieran Flanagan, HubSpot's CMO, investing as a Sequoia scout.

Founded by Andreas Kullberg, Agaton builds agentic AI software that analyzes customer conversations in real time. The platform detects buying intent, sentiment shifts, and churn risk, then translates those signals into recommended actions for sales and service teams. It's not transcription. It's not call recording with keyword alerts. It's something closer to giving every sales rep an analyst who listens to every interaction and tells them what to do next.

"The companies winning today aren't those with the biggest sales teams," Kullberg says. "They're those with the smartest ones, augmented by AI that transforms every customer interaction into strategic intelligence."

That's founder-speak, sure. But the numbers behind it are hard to dismiss. Inception Fund partner Erik Lindblad noted that Agaton signed "millions in ARR within a year, while maintaining fanatical customer focus." For a company operating in stealth, that level of commercial traction before a public launch is unusual. It suggests either an exceptionally sticky product or an exceptionally connected founding team. Probably both.

The conversation intelligence market isn't new. Gong, Chorus (now part of ZoomInfo), and Clari have been mining sales calls for insights for years. But Agaton's pitch is different in a specific way: it's agentic. Rather than generating dashboards and reports that someone has to interpret, the system produces recommended next actions. The AI doesn't just tell you that a customer sounded frustrated. It tells the rep what to say next time.

A $10 million seed is large by European standards, where the median seed round still hovers around $2-3 million. Agaton's ability to command that figure suggests investors are pricing in the existing revenue, not just the vision. The angel list reinforces the thesis: Peter Sarlin built and sold one of Finland's most successful AI companies. Kieran Flanagan sits at the intersection of marketing and sales technology at one of the world's largest SaaS companies. These aren't passive checks.

Most conversation intelligence tools work retrospectively. They record calls, transcribe them, highlight keywords, and generate summaries. Managers review the summaries. Maybe something changes. The feedback loop is measured in days or weeks.

Agaton's real-time approach collapses that loop. If a customer's sentiment shifts during a call, the system detects it and can surface guidance while the conversation is still happening. If buying signals emerge across multiple interactions with a single account, the system aggregates them and flags the opportunity. The value proposition isn't better reporting. It's faster action.

Whether that's actually possible at scale, consistently, across different industries and communication styles, remains to be proven. Real-time NLP is computationally expensive and culturally sensitive. Sarcasm, industry jargon, and multilingual conversations all create noise. Agaton will need to demonstrate that its AI handles these edge cases reliably enough for enterprise buyers to trust it with their revenue operations.

There's something interesting happening in Stockholm's AI scene. While much of the AI startup world is raising on vibes and demo videos, a subset of Swedish companies keep launching with revenue. Lovable hit $400M ARR with 146 people. Agaton came out of stealth with ARR already accumulating.

Part of this is cultural. Swedish founders tend to build quietly. Part of it is the investor ecosystem: Nordic VCs have long preferred traction over hype, and the angel networks in Stockholm are dense enough that a good product finds early customers through warm introductions rather than press coverage.

Agaton's next challenge is scaling beyond its initial customer base without losing the product focus that got it here. Enterprise sales cycles are long, and the competitive field includes well-funded incumbents. But starting with revenue instead of a pitch deck is a luxury most seed-stage companies don't have. Agaton seems intent on using that head start wisely.